Latest / The Innovation Show / Carliss Baldwin & Kim Clark — Design Rules: How IBM Created Ecosystems (and Competitors)

Transcript

Carliss Baldwin & Kim Clark — Design Rules: How IBM Created Ecosystems (and Competitors) · The Innovation Show

• 0:00
  today's book is a prerequisite text for anyone thinking about platforms,
• 0:05
  ecosystems, open architecture, or deep structure of disruption.
• 0:09
  On the innovation show, we have spent years interviewing people on how
• 0:13
  industries break apart and reform, from Christensen to Osterwalder,
• 0:18
  from McGrath to von Hippel.
• 0:20
  Today's book is the Rosetta Stone that connects all of them.
• 0:24
  Once you know what a modular operator is and how it generates
• 0:28
  real options, you can reread so many other thinkers at a deeper level.
• 0:33
  Three lenses for us to apply today.
• 0:36
  One, for founders or operators, modularity is the engineering
• 0:41
  precondition for ecosystem strategy.
• 0:43
  You cannot platform an interdependent product.
• 0:47
  Two, investors.
• 0:49
  The book formalizes why modular industries spawn many small, valuable firms, while
• 0:56
  integral industries spawn few large ones.
• 0:59
  It will help you know what to look for and what questions to ask.
• 1:03
  And for incumbents, once a competitor splits your architecture, you face
• 1:08
  an irreversible value migration.
• 1:12
  The IBM System/360 story is the cautionary tale that we might share today, and
• 1:18
  that indeed is the book's tragic irony.
• 1:21
  The act of modularizing your product is the very act of inviting your
• 1:26
  future competitors into existence.
• 1:29
  We know well the innovator's dilemma.
• 1:31
  This is the mechanic that underpins that concept, and it is e- even
• 1:36
  more important in this age of AI.
• 1:39
  It's an immense honor to host both authors in an innovation show exclusive
• 1:45
  of their, in my humble opinion, massively underrated, underappreciated book,
• 1:51
  probably because it was too early, and now it is certainly of its time.
• 1:56
  First up, welcome back to the show, friend of the show, and a guy I call a friend,
• 2:02
  Kim Clark, you're very welcome back.
• 2:04
  It's good to be here.
• 2:05
  Thank you
• 2:06
  Great to have you back always, Kim.
• 2:08
  I always enjoy our chats.
• 2:09
  And then a lady I have just met and I hope to have back many more times on
• 2:14
  the show if I pass my audition today, Carliss Baldwin, welcome to the show
• 2:19
  Thank you so much, Aidan.
• 2:20
  It's a great honor to be here and wonderful to see Kim
• 2:23
  the honor is all mine.
• 2:25
  I thought we'd start because there's an entire vocabulary for the book.
• 2:29
  It took me a while to learn it.
• 2:31
  I was r- I had written it down in one column when I was reading the
• 2:34
  book, and we don't have time to share all that vocabulary, but I thought
• 2:39
  we'd start with the key term itself.
• 2:41
  Carliss, maybe ladies first, and then Kim can build on it.
• 2:45
  But the term itself, what is a design rule?
• 2:48
  A design rule is a formal constraint made explicit that allows people
• 2:59
  are not directly involved in the creation of the design to create
• 3:06
  complementary designs are compatible.
• 3:11
  So they, they don't have to be there at the creation.
• 3:16
  They are not integrated into the original process.
• 3:22
  But with the design rules in place, can build a platform system.
• 3:30
  Design rules are, in fact, the platform.
• 3:33
  A platform system that can be extensible and creates many, many,
• 3:39
  many options develop in different ways
• 3:43
  Kim, maybe you'll come in and give us an example of a design rule at play
• 3:48
  . Clothes are a great example of, you know, you can buy, you can buy , shoes
• 3:54
  and pants, shirts, jackets, dresses from completely different firms that
• 4:02
  have no idea the other people are working on stuff, and it all fits
• 4:07
  together when you do it because there are a set of design rules.
• 4:11
  And if you start, if you start looking for them, they turn out to be everywhere.
• 4:19
  They are a really important feature of modern life.
• 4:23
  Without them, we, we would not have what we think of as, you know, everyday things.
• 4:30
  You think of… I mean, just, just walk into your kitchen sometime
• 4:34
  and look around you and you see all-- You see modularity in action.
• 4:39
  you know, the frying pans are not 18 inches wide.
• 4:45
  They're just not.
• 4:46
  The burners on your stove, they're all kind of about the same size, and they
• 4:52
  all plug into the wall in kind of the same way, and they use electricity
• 4:55
  the same way, and they plug into the gas, and everybody can work together.
• 5:00
  You look at, you know, you just walk around your house and there's modularities
• 5:03
  everywhere because we have a set of design rules that allow completely independent
• 5:09
  firms who were not involved in designing the rules to function and operate.
• 5:14
  Where it becomes really powerful is in a technology that where
• 5:21
  modularity unlocks incredible value.
• 5:26
  And that's what we, that's what we discovered really when
• 5:30
  we dug in deeply into the IBM story and the computer industry.
• 5:35
  Because inherently in that technology, once you got, once you got away from
• 5:41
  analog devices and got to digital, you were moving into a technology that was…
• 5:48
  It's so incredibly powerful that it's unfolded the way it has over
• 5:54
  the last, you know, almost now 60 years because of modularity.
• 6:00
  I mean, we, we love to say that modularity was the underpinning of Moore's Law.
• 6:05
  I mean, you don't get Moore's Law without modularity, and you don't get,
• 6:10
  you don't get everything we see around us that's electronic and digital.
• 6:14
  all depends on, on modularity.
• 6:17
  Maybe we'll build one term in there because the, you say in the book,
• 6:21
  "Design is the process of inventing objects or things," like you said,
• 6:26
  "that perform specific functions.
• 6:28
  But these options, the products of human intelligence and effort, are then
• 6:32
  called artifacts." And artifact is a key term to understand modularity as well.
• 6:37
  Maybe we'll explain that.
• 6:38
  you have to go to Carliss to do that because she is, she's the wizard behind
• 6:44
  turning anything into an artifact.
• 6:46
  And,
• 6:47
  Oh, no, no, no, no, no.
• 6:49
  No,
• 6:49
  my
• 6:50
  technologies have of a fundamental nature, and before the digital, before digital
• 6:58
  technology almost all advanced technology was various forms of what do you call it?
• 7:07
  Flow processes.
• 7:08
  I mean, the idea was you made things go-- be bigger if they-- and they went faster,
• 7:14
  and as you sought, sought and speed companies also get bigger and faster.
• 7:22
  And that's a fundamental flow process.
• 7:27
  So it, it's A is connected to B is connected to C, and, and to a
• 7:33
  first approximation, everything happens in order, and there's a set
• 7:38
  of prescriptions for making those, those technologies more efficient.
• 7:46
  It's, it's you, you standardize the steps, you them, you make
• 7:53
  each step the same length, you keep looking for the bottlenecks.
• 7:59
  That was that was production as I was taught when I was an MBA student.
• 8:06
  K-Kim skipped the MBA.
• 8:08
  He, he went straight to the PhD. But there's a whole, you know, a set of
• 8:15
  technologies from railroads to steel to for-- to automobiles to airplanes.
• 8:23
  They all, they all s-fundamentally work the same way, and they do not
• 8:27
  really reward modularity, although sometimes you can modularize a flow
• 8:32
  process and run things in parallel.
• 8:35
  But to, to first approximation, modularity is not a goal.
• 8:42
  So digital technology comes along, and it has a very, very
• 8:47
  interesting structure in that of it is hardware, and that's circuits.
• 8:56
  And circuits basically are not too different flow processes.
• 9:01
  It's, it's electrons are flowing through the circuit.
• 9:06
  But the other half is software, and they are strict complements.
• 9:12
  You-- Without one, the other is worthless.
• 9:14
  So, S- so you have this immediate dilemma because software then, this
• 9:20
  is pre-Claude soft- software then was a, was a, it w- it was a writing
• 9:27
  process, a handwriting process.
• 9:30
  And once it was written, you didn't wanna rewrite it 'cause
• 9:34
  that was very, very arduous.
• 9:37
  So here you are, a manager, and you've got one thing that has Moore's law
• 9:41
  going for it, where you just wanna turn over the designs as fast as you can
• 9:47
  and get faster and faster and faster.
• 9:49
  as the interface of that and the instructions is this enormous barrier
• 9:57
  where in order to run a different, faster set of circuits, you have to
• 10:04
  rewrite the instructions scratch by hand.
• 10:10
  And this was something that IBM was all geared up to come out with
• 10:17
  model after model after model, and they wanted a whole-- Well, the,
• 10:20
  in the-- Originally, they didn't.
• 10:22
  They just wanted to do upgrades.
• 10:24
  They didn't think of a whole product line.
• 10:26
  But here they were stuck with this software problem, and their, their
• 10:30
  customers did not wanna rewrite their hard-won software buy the
• 10:36
  next bigger, better machine.
• 10:40
  So this was like a, a A tremendous problem for sales.
• 10:47
  Was just gonna stop it in its tracks.
• 10:49
  No- nobody wanted to do upgrades.
• 10:52
  And so IBM's chief executive gathered together 12 of her 14 engineers
• 11:03
  with different but complementary
• 11:11
  I used to say they put them in a motel in Cos Cob, Connecticut.
• 11:14
  Turns out it wasn't Cos Cob, it was a, a nearby village in Connecticut.
• 11:19
  They did this in late November, and they said, "You can't come out
• 11:22
  until you solve this problem." I mean, Christmas is coming, right?
• 11:27
  So so they-- everybody understood the problem, and they solved it creating
• 11:35
  design rules, which said that if you follow these rules that are at the top of
• 11:41
  the heap, then your software be performed the next largest machine the next largest.
• 11:51
  So, so, so you never, you never have to rewrite the software.
• 11:55
  It just gets run on a bigger machine, and that bigger machine can also
• 12:00
  run more software, bigger software.
• 12:04
  And so it, it ended up being quite a brilliant suggestion.
• 12:10
  And they understood, the, the, the guys who, who were part of this
• 12:15
  group, they said it just made sense.
• 12:18
  By the end of the… Took them-- gave the report, I think, on December 27th, but I
• 12:24
  think they got to go home for Christmas.
• 12:27
  They said, "We became more and more convinced that it would
• 12:30
  work," and indeed it did.
• 12:33
  And that was the first real conscious exercise of a modular platform strategy.
• 12:42
  We actually have lots of plat-- The high street of a village is a platform.
• 12:49
  It's got all these stores, they're all different from each other they, they,
• 12:54
  they benefit from the presence of others.
• 12:57
  But nobody thought of that as a strategy.
• 13:00
  That was just a street.
• 13:02
  But with IBM and the 360, they needed something formal and
• 13:10
  transferable and a set of principles.
• 13:14
  it was-- I, I, I claim it is the one of organization design that was
• 13:23
  not just A natural extension of what was happening, but was invented.
• 13:29
  It was an invention.
• 13:31
  So, so flow processes, they're just, you know, they go back to the pyramids.
• 13:37
  We've always had sort of these people agree that it's a platform
• 13:43
  and they use its flexibility.
• 13:46
  Job shops are just the way people get work done if they don't have a lot of work.
• 13:50
  You know, if it's a, if it's a small enterprise and they need
• 13:54
  flexibility and variability.
• 13:57
  A line staff organization like an army or a government or a corporat-
• 14:03
  big corporation, that's just sort of you start with job shops and then
• 14:07
  you link them together with staff.
• 14:09
  Or you have flow processes and you link them together with staff.
• 14:14
  But modularity, you have to actually set out to make it modular by creating the
• 14:19
  design rules and having the modules obey.
• 14:22
  So Aidan, let me, let me go back to the idea of an artifact, and I'm gonna
• 14:28
  use the one that we used in the book.
• 14:30
  See this?
• 14:33
  It's a mug, right?
• 14:36
  And, it has certain elements to it.
• 14:39
  You can see them now.
• 14:41
  And this is something that we believed in very strongly.
• 14:45
  You can see it all through the book, and all through Carliss's subsequent
• 14:49
  work, which is, it really pays to know the details of the design, to
• 14:57
  actually understand in, in pretty pretty significant detail what is
• 15:02
  involved in creating a design, right?
• 15:07
  So, you know, you've got a handle, you've got a receptacle, you've got a lid.
• 15:12
  But if you-- If you've read it, I mean, there's a lot that goes into like, well,
• 15:17
  how does the lid fit on this thing?
• 15:18
  How does it, you know, how do you kinda make sure that it's right?
• 15:22
  And, and you can see immediately as you design this thing that there are
• 15:27
  all sorts of decisions you have to make about the structure of this, "artifact,"
• 15:33
  which is essentially a collection of little design choices that you make
• 15:39
  in order to create something that someone can use and use effectively.
• 15:45
  So if you think about that idea and go back to the computers before the IBM
• 15:51
  360, they had the same exact issue.
• 15:56
  And-- But the only thing they knew and understood how to do was to, was to,
• 16:05
  create the hardware, which required a lot of design choices, because you had to
• 16:10
  create interfaces just like you do here.
• 16:12
  Like, what's the interface between this and the customer?
• 16:16
  Well, right now it's a straw, okay?
• 16:19
  But you have other kinds.
• 16:20
  You could make this like, you know, like a little flippy thing
• 16:24
  that you flip up and, you know.
• 16:26
  So you had choices about how to design the, the, the hardware of
• 16:31
  the computer, and then you had to figure out how to make it do things.
• 16:36
  Well, the problem they had was that the only way they knew how to make the
• 16:42
  computer actually work was to create a set of instructions for the computer.
• 16:50
  in the early days, they didn't know how to write software.
• 16:56
  fact, the idea of software wasn't even around.
• 16:59
  The only thing they knew how to do was to write instructions into the the
• 17:05
  hardware, and then it's called, it's actually called the instruction set,
• 17:10
  and it gets hard-coded in circuits.
• 17:13
  They knew how to do that, and then they knew how to tell the, the customer
• 17:18
  how to access those instructions.
• 17:21
  But it was all… In a sense, it was just like this artifact, you know.
• 17:26
  You designed it, it all fit together, you figured out how to make it work.
• 17:29
  It's fantastic.
• 17:30
  Okay?
• 17:32
  The problem was that they wanted to create new hardware in this car.
• 17:37
  And so that, that all unfolded.
• 17:39
  And, and the artifact itself, the hardware in the computer, fundamentally
• 17:46
  didn't, change all that much.
• 17:49
  It just got faster and bigger, so you could do more with it.
• 17:55
  So that meant you could draw-- could, you could make the, you could make the
• 18:00
  little circuits smaller, essentially.
• 18:03
  And so the electrons didn't have to travel as far.
• 18:06
  And essentially, that's what's been going on for on, you know, it's Moore's law.
• 18:10
  So that artifact, the, the fundamental issue was that uh, it was as if in order
• 18:18
  to, to have a mug, now had to create some kind of instructions that the customer
• 18:25
  had to give the mug before it would work.
• 18:30
  You know?
• 18:30
  It is true
• 18:31
  the next thing that happened was that they figured out that, oh All these different
• 18:40
  things need that customers really like.
• 18:43
  Like, for example, they liked to be able to print out what they were
• 18:49
  working on create a document, right?
• 18:53
  printers became a thing.
• 18:56
  oh, we can actually now, because guys in-invented another piece of hardware,
• 19:02
  and I was in the plant in San Jose, and I saw these things, these enormous disks.
• 19:09
  These were like-- They weren't like little ones.
• 19:12
  These were great big, huge disks.
• 19:14
  And they figured out how to make a device where you could send a digital signal
• 19:20
  to the device, and it would encode it in the hardware that is in the, in the disk.
• 19:26
  And they could put them in this big stack, and you could sell a disk drive, you could
• 19:32
  store enormous amounts of information.
• 19:33
  And the customer said it was fantastic, right?
• 19:37
  those devices existed before the 360, but they were all
• 19:42
  linked to that instruction set.
• 19:45
  And so only the software they had to change.
• 19:48
  All of a sudden, the customer had to buy new printers and new disk drives, and it
• 19:54
  was a mess, and they didn't wanna do it.
• 19:58
  And, And, so they figured out.
• 20:00
  But as you mentioned earlier uh, once IBM split up and they figured out how to do
• 20:07
  it… By the way, doing it was not easy.
• 20:11
  We can talk about it like, "Oh, they figured out modularity. they just decided
• 20:14
  to make it modular." Well, that was, like, super difficult, really hard to do.
• 20:20
  Uh, but they figured out how to do.
• 20:22
  it, they pulled the instruction set out of the hardware, and they had to figure
• 20:27
  out how, how do you make the machine talk to the software and so forth, and
• 20:31
  they figured out how to do it, right?
• 20:33
  So they split it, and now you could upgrade the machine, and you
• 20:37
  didn't have to change the software.
• 20:38
  And they figured out, well, if you have design rules for that, now
• 20:41
  you can have design rule for the printers, rules for the disk drives.
• 20:44
  You don't have to buy new disk drives.
• 20:47
  By the way, we have new disk drives if you want them, but you don't have
• 20:50
  to buy them because we can upgrade your computer and give you a whole new
• 20:53
  thing, and it's faster and so forth.
• 20:55
  Well, that was the beginning of the fundamental change in
• 21:01
  the computer industry, right?
• 21:04
  Uh, it, and it, it was-- You could see it in action.
• 21:08
  You could watch it happen because 12 engineers that worked in San
• 21:12
  Jose, where they made all their disk drives, walked out the door.
• 21:16
  And they formed, uh, new companies, uh, started with new-- one new
• 21:20
  company to build disk drives.
• 21:23
  And that spawned this enormous disk drive industry, and they became really much,
• 21:29
  much faster than IBM and much better at building drives, even though IBM had the
• 21:34
  world's best scientists working on it.
• 21:36
  The guys… It turned out that once you figured out, oh, it's like how
• 21:40
  to encode in the drive, you can make them smaller and faster and so forth.
• 21:45
  Anyway, and that spawned the disk drive industry, and the same thing
• 21:48
  happened with printers and every other peripheral that IBM had.
• 21:53
  But IBM figured out, "Oh, this is actually really good for us." So they
• 22:02
  didn't try to kill these people because what they just realized is, "Oh, this
• 22:08
  means we can create an IBM compatible," essentially, e- no, we had those
• 22:15
  language in those days, ecosystem, because the 360 became a platform.
• 22:20
  And, uh, somebody once said the, uh, after the 360… This is hard to
• 22:26
  imagine today, but the announcement of the 360 was, like, an enormous event.
• 22:33
  It was covered by all the newspapers.
• 22:36
  It was covered by TV.
• 22:38
  enormous event in New York.
• 22:41
  And the 360 got introduced people realized, "Oh, this is completely
• 22:44
  different." A few years later, not very many years, people talking about IBM
• 22:52
  as a competitor, and they just said, "IBM is the environment." And, and if
• 22:59
  you look at the graphs we'd created, Carliss in her Magic created all these
• 23:03
  beautiful graphs about value and stuff.
• 23:06
  I think, if I remember right, IBM had something like, I don't
• 23:10
  know, 90% of all the value in the computer industry was in, in IBM.
• 23:15
  And that was with all this disk drive and printers and
• 23:18
  all the stuff that was around.
• 23:20
  Yeah, there you go.
• 23:21
  And it was, like, unbelievable, you know?
• 23:26
  It was incredibly valuable.
• 23:29
  So that was the, uh, that was the peak you saw there.
• 23:34
  The peak.
• 23:35
  It's good
• 23:36
  there's so much here.
• 23:37
  I, I mentioned it's the Rosetta Stone, and what I mean there is when you
• 23:42
  read it, it's, it took me a while now, I have to say, and I see the
• 23:46
  depth of research that you've done.
• 23:48
  And Carliss, now you've gone and released Design Rules 2 which
• 23:52
  is just a massive body of work.
• 23:54
  It's a lifetime of work.
• 23:56
  But when you, when you start to get it, you start to go, "Well, now I
• 24:00
  understand actually the source of where Clay Christensen, may he rest in
• 24:06
  peace, got the idea of The Innovator's Dilemma," because it's the same thing.
• 24:09
  It's like, by doing this is a dilemma.
• 24:12
  I have to do this because I can't continue to grow.
• 24:16
  I can't continue to work this way.
• 24:18
  I need to start to let go of parts of it.
• 24:21
  I need to create an ecosystem like IBM did.
• 24:24
  But if I do that, I can create competitors, and the competitors
• 24:27
  can actually be better than at some parts of the thing that I used to do.
• 24:31
  And it, and it's a re- it's a huge dilemma.
• 24:33
  And, and as you told me before, Kim, that the IBM, they, they did try to sue
• 24:38
  some of the guys who walked out the ru- who knew the design rules, 'cause the,
• 24:42
  the design rules were almost like the blueprint maybe we'll talk about that
• 24:46
  Because I think for, for people, when they listen to this or they hear or they
• 24:50
  see the book, they'll be f- afraid of it
• 24:53
  like I was.
• 24:55
  But w- but when you do start to s- understand some of these rules,
• 24:59
  you just have these ah moments.
• 25:01
  And, and one of those ah moments for me was understanding, and I'm
• 25:05
  pretty sure Steve Jobs read your book and went, "We're gonna do that.
• 25:12
  We're gonna understand the design rules, except we're gonna
• 25:14
  orchestrate the entire ecosystem,
• 25:17
  Yeah, you look at that diagram that you've
• 25:18
  got on the screen, right?
• 25:20
  And, uh, peak years are late '60s up to about 1970.
• 25:26
  And it just was, it just was the
• 25:28
  case that some
• 25:29
  people figured out, "Oh, you know, can create computers that are not as
• 25:40
  big, not as functional, they really provide uh, a, a different service
• 25:48
  to people." And that's where, you know, Digital Equipment Corporation
• 25:54
  came from right about this time.
• 25:57
  And, , they understood that you could, you could now start building computers, and
• 26:04
  you didn't have to do everything yourself.
• 26:07
  And that's, that was an amazing moment.
• 26:10
  There is something that's really important, though, when you talked
• 26:12
  about competitors and understanding things about if I, if I'm a-- if I'm
• 26:19
  involved in creating the platform, it's enormous value that's created.
• 26:24
  How do I capture it?
• 26:27
  And one of the things that we can see in the history of this industry is that
• 26:32
  in order to do that have to understand what's going on a depth that will
• 26:39
  really allow you to make good decisions.
• 26:42
  Like I'll, I'll just give you two examples at IBM.
• 26:48
  One was that IBM owned 30% of Intel at one time
• 26:59
  and it's because Intel was having trouble and was under a lot of pressure
• 27:05
  s-switching from being a memory company to being a processor company.
• 27:09
  IBM is using Intel's processors.
• 27:12
  So right away, you are now deep into the platform, the ecosystem, because
• 27:19
  there are companies now making processors that only IBM used to make.
• 27:26
  And now IBM stepped in because they were using Intel's processors, and they
• 27:33
  invested and, and own 30% of Intel.
• 27:37
  They also, at about the same time, decided-- and this is when they
• 27:42
  decided to go into a, to a small computer and personal computer They,
• 27:49
  uh, contracted with Microsoft to build the operating system, right?
• 27:56
  But they didn't, which they could have easily done, because
• 28:00
  Microsoft was a tiny company.
• 28:02
  They did not put in the contract who owns the operating system.
• 28:09
  And IBM could have easily owned the operating system, easily.
• 28:14
  And of course, that-- those two things, the processor and the operating system,
• 28:20
  became the only sources of profit in the, in that ecosystem that evolved over time.
• 28:29
  The only people who made a lot of money were the people who owned the
• 28:31
  processors and the, and the operating system, until, of course, other people
• 28:36
  figured out how to do the processors, and other people figured out how to
• 28:42
  generate operating systems, right?
• 28:43
  Apple being one.
• 28:45
  So it, it's-- I have a-- The two of us have a colleague at HBS
• 28:52
  who's um, he's one of my mentors.
• 28:55
  His name's Bob Hayes.
• 28:57
  And Bob once, wrote a paper, wrote a… It was a case, I think,
• 29:01
  about the Swiss watch industry.
• 29:05
  it, covers like hundred years.
• 29:09
  And he said-- And he taught me this.
• 29:11
  He said-- remember, the game is never incontestably won."
• 29:22
  It's never incontestably won.
• 29:24
  There's always people who will come in.
• 29:27
  So it looks like, at certain points in time, it looks like this
• 29:31
  company just has, like IBM, has a clo- total lock, which it did.
• 29:37
  But only lasted… Look at your diagram.
• 29:40
  It only lasted for a few years because, you know, once they, the technology got…
• 29:46
  Now, they could've, they could've, if they'd understood the platform and
• 29:51
  understood modularity, they could have figured out where's the real value
• 29:55
  that's gonna be really, really hard to copy, and then owned it, right?
• 30:00
  But they didn't do that
• 30:02
  - I find that fascinating where there's, uh, many case studies like that
• 30:06
  where they had all the ingredients, but they didn't have the lens.
• 30:10
  And for me, the book is a lens.
• 30:12
  It's a new lens to look at things, and that's why I said about that
• 30:15
  intro, if you're an investor or you're a startup founder, you can
• 30:19
  start to see things differently, and particularly at this moment in time.
• 30:23
  Carliss, maybe you'll say a word on this moment in time.
• 30:26
  I mentioned that the book's of its time, and it's, for me, what AI
• 30:30
  seems to do is, is modularize , it explodes things into pieces.
• 30:36
  And then if you know what the Lego bricks are and you can reconstruct
• 30:40
  them, you can create something totally different and totally new value
• 30:44
  I'm glad you brought up AI.
• 30:45
  It's, it's something that is, is quite puzzling
• 30:53
  in the large language models
• 30:57
  one of the things in days with Kim, I kept saying, "How do we know it's
• 31:02
  modular? How, how do you…" You know, we-- everybody talks about modularity.
• 31:08
  Sun Microsystems was our first case.
• 31:12
  How do §you know?
• 31:14
  And
• 31:17
  the thinking was very fuzzy at the time.
• 31:21
  kinda knew because there were little boxes and you could hook stuff together.
• 31:28
  we came across a technology called design structure matrices, which actually you to
• 31:37
  identify the tasks being done by people or machines, the transfers of energy
• 31:47
  and information those tasks that allow the task to be done in the right order.
• 31:54
  And
• 31:57
  I was just entranced.
• 31:58
  I'm still on a campaign.
• 32:00
  I call them-- They were called DSMs, I call them TSMs now.
• 32:04
  But I'm, I'm on a campaign to get people to look at the
• 32:08
  bloody TSMs of their processes.
• 32:13
  Kim said, "That's the geekiest thing I've ever seen." And I
• 32:18
  said, "Kim, you don't understand.
• 32:20
  It unambiguously
• 32:22
  tells
• 32:22
  you if you have a modular system.
• 32:25
  You don't have to believe the, you know, the, the, the necromancers, the, the
• 32:31
  wizards saying, 'Oh yeah, it's modular.
• 32:33
  We use modularity.' the, the little, the little transfer dots
• 32:39
  will tell you." and you know
• 32:46
  it makes the architecture of the what's going on visible and understandable.
• 32:56
  And Kim's first paper on innovation was with Rebecca Henderson, and it
• 33:04
  was about architectural innovation and the failure of established firms.
• 33:10
  a very, very, famous, very influential paper.
• 33:14
  And what it boils down to, and he didn't use TSMs, they had a wonderful
• 33:21
  case study of I think photolithography.
• 33:25
  When you have architectural innovation, and this speaks to evolution in
• 33:31
  the computer and other industries, changing the configuration of the
• 33:37
  tasks and transfers inside the module.
• 33:41
  And companies often can't cross that chasm.
• 33:46
  They can't go from one kind of class-- Many times, 'cause they don't know really
• 33:54
  going on inside their architecture.
• 33:58
  And, so when it comes time to, make a change, they have no,
• 34:06
  no insight into the new thing.
• 34:08
  A, a good example is Amazon.
• 34:12
  You know, be- before Amazon, there was Sears-- there, there was Walmart,
• 34:17
  and before Walmart, there was Sears.
• 34:21
  Sears couldn't turn into Walmart, and Walmart couldn't turn into Amazon.
• 34:27
  if you think about it, what Amazon did was it has-- it, it had a, a module that sold
• 34:35
  things, and it replaced stores a website.
• 34:42
  And then it had a module that fulfilled things, that was logistics.
• 34:47
  And what Bezos did within a year of founding Amazon is he hired a
• 34:53
  bunch of the second from the top guys from Walmart
• 34:58
  and brought them to Amazon to set up a f- a fulfillment process.
• 35:04
  And the ful- the architecture of the fulfillment process was a
• 35:08
  standard logistics.
• 35:09
  I
• 35:10
  mean,
• 35:10
  they, they are very admirable systems, but it was something that was well
• 35:15
  understood in that Sears-Walmart line.
• 35:19
  but going from stores to a website,
• 35:23
  how would you do that?
• 35:25
  If, if,
• 35:26
  you
• 35:26
  know,
• 35:26
  i-i-if-- particularly if, you have a base,
• 35:29
  an embedded base
• 35:32
  managers and trained employees
• 35:35
  and, and,
• 35:36
  you know, inventory processes and
• 35:39
  you know,
• 35:39
  customers who love nothing more than a bookstore.
• 35:46
  you just--
• 35:46
  it's, it's
• 35:47
  it's like an impossible
• 35:49
  dream.
• 35:50
  and that was
• 35:51
  what Kim and Rebecca
• 35:52
  pointed out
• 35:54
  in, as I said,
• 35:56
  not, not, his
• 35:57
  first paper, but his first paper on innovation, I Think.
• 36:02
  yeah.
• 36:04
  It's your, it's
• 36:05
  Carliss
• 36:05
  stuff.
• 36:06
  uh, these, little off X's in a design structure matrix indicate
• 36:13
  inter-- a, a point of interaction where you can't do this until you do
• 36:16
  this, or this needs to be connected to this in some way through either
• 36:20
  information or material or whatever.
• 36:24
  But underneath that, in an are people and processes sometimes hidden
• 36:33
  rules of thumb and other things that allow that thing to function.
• 36:39
  The thing that we did was based, the, the we did, and it, it was Rebecca
• 36:44
  Henderson's thesis, um, which was a, remarkable thesis in and of itself.
• 36:50
  But it
• 36:53
  it was organizations that were small
• 36:58
  These were uh, had 20 engineers.
• 37:04
  They weren't huge the photolithography industry.
• 37:10
  When I talk about organizations, I'm talking about those that were in the
• 37:12
  photolithography business, 'cause some of them were uh, divisions of big companies.
• 37:17
  But even within a small group, this process works, where the change to-- in
• 37:24
  the architecture becomes very difficult to transition because of the organizational
• 37:30
  communication and patterns of work and design rules that been in place before.
• 37:37
  But I want to comment on Walmart and Amazon because it's a really
• 37:43
  great, great example, which is uh, Amazon was introduced in about '94, I
• 37:53
  think.
• 37:55
  Five.
• 37:56
  So we're, we're 20 what?
• 37:58
  31, 31 years since then.
• 38:02
  Today
• 38:05
  Walmart competes really successfully with Amazon in online delivery and
• 38:11
  running a website and so forth.
• 38:14
  How-- But Walmart lived to make it because they were base-- They were--
• 38:20
  They had this huge base business, right?
• 38:23
  And it took a long time for the customers to kind of figure out,
• 38:29
  "Oh, I can get stuff from Walmart."
• 38:31
  It took a while.
• 38:32
  And even today, online retail is not huge factor in the retail industry.
• 38:39
  It's like 20%, something like that.
• 38:44
  People like to go to stores.
• 38:45
  They like to shop, right?
• 38:48
  Walmart had enough time and, and enough stickiness in the customers
• 38:54
  that they could figure out, and it took them a, it took them a while.
• 38:59
  Um, wrote a case about a young man who, his name's Sam Bernards.
• 39:04
  He went after he, he was a uh, undergraduate major in, uh, in
• 39:12
  computer, uh… I guess he was a-- It was in electrical engineering.
• 39:17
  Anyway, he did an MBA.
• 39:19
  He went to Walmart because they hired him to help them figure out how to go online.
• 39:25
  This would've been 20 years ago.
• 39:28
  Okay?
• 39:29
  Uh, and maybe less, Anyway, but he built-- He, he got them online.
• 39:39
  He built a team, about 80 people.
• 39:41
  They g- they built a website.
• 39:43
  They've invented how to take Walmart… Walmart knew a lot about
• 39:46
  logistics, how to do that, build an operation that could be fulfillment.
• 39:52
  And then it became like a $300 million business, which was tiny in Walmart.
• 40:02
  Walmart had the time and the resources to live while they
• 40:07
  figured out how to make transition.
• 40:09
  The guys in photolithography, the incumbents who faced an architectural
• 40:15
  change, and there were about, Rebecca documented there were about six of
• 40:19
  them that she documented, because they happen every couple of years.
• 40:23
  The incumbents invariably didn't just have trouble transitioning, died
• 40:32
  Because customers could switch immediately to the new technology.
• 40:39
  They saw it was better, it was faster, it was cheaper, it was-- had higher
• 40:44
  quality, could do finer line widths.
• 40:47
  And the old stuff, the guys just had a hard time, and once they
• 40:50
  had a hard time, they were dead.
• 40:53
  So the dynamics of the industry also depend on the customers and what's
• 40:59
  going on in the customers' processes.
• 41:02
  And I wrote another paper about that, actually.
• 41:05
  Uh, And you know, and the customer-- Sometimes customers get stuck, like the
• 41:11
  people who like to go to retail stores.
• 41:13
  They like to go to retail stores.
• 41:14
  It, it's sort of… It's not just the m- stuff they want, but they
• 41:18
  want the experience of choosing and walking around and doing stuff.
• 41:23
  I'm married to a person that loves to do that.
• 41:25
  Me, I like guy shopping, which is I need X, I go into the store, X, and I come out.
• 41:32
  Why shouldn't I just go online, I go I want
• 41:34
  "X, send me X," and they send it to me tomorrow.
• 41:36
  I say, "Great."
• 41:38
  But I have a spouse who loves to shop.
• 41:42
  In fact, I'll tell you a funny story.
• 41:43
  You can probably edit this out, but it's a funny story.
• 41:47
  When my twins, I have turned… Let's see, how old were they?
• 41:50
  I think they were about 16.
• 41:51
  When they turned 16, I took them with, with Sue, my wife, to New York City
• 41:56
  for a Broadway show and shopping.
• 41:58
  So we got up from-- and went shopping, and the first place was, was The Gap.
• 42:04
  It's this huge, beautiful store in, on, on uh, in New York.
• 42:08
  And they had a whole lounge area for guys.
• 42:12
  So I sat in the lounge area for guys, and my daughters and my wife went shopping.
• 42:18
  And they weren't gone for more than like a half hour, and
• 42:21
  they came back with nothing.
• 42:23
  I said, ""What happened?" And they said, "There are no bargains in this store."
• 42:31
  And
• 42:32
  Oh, jeez.
• 42:34
  taught me
• 42:35
  that for them, shopping was like hunting.
• 42:41
  And it didn't make any sense to them to fa- pay full price for anything.
• 42:46
  they were after bargains.
• 42:49
  And we spent a whole day walking up and down stores in New York, and they never
• 42:53
  bought a thing till we got to Chinatown.
• 42:57
  And in Chinatown.
• 42:58
  there were all sorts of bargains and, you know.
• 43:03
  so
• 43:03
  some-- if you have customers that are so wedded to shopping as a thing,
• 43:08
  you can live and survive, 'cause it takes a long time for that to happen.
• 43:14
  Customers who are really focused on speed and line width and, you know,
• 43:20
  it, it, it takes, it takes like 30 seconds for them to switch because the
• 43:26
  equipment
• 43:27
  works the same way.
• 43:28
  It just works differently, and they figure it out.
• 43:31
  easy to do, and the incumbent is dead
• 43:37
  So you have, you have, to understand the dynamics of the technology.
• 43:45
  You have to understand what modularity really allows you to do.
• 43:49
  You have to understand your customers and their processes and how it works
• 43:53
  in
• 43:53
  order to figure out where's the value being created, how does it work, so forth.
• 43:59
  And, you know, a lot of the book was-- grew out of our understanding
• 44:03
  of, of real options, right?
• 44:05
  That a sudden modularity creates real options.
• 44:09
  And then we figured out, oh, people can design stuff in such a way so as to create
• 44:15
  different kinds of modularity, different kinds of stuff that they can use.
• 44:22
  And if you understand that, the value of that… So you have a certain
• 44:27
  amount of value that's just modular.
• 44:29
  Like ma-t- making something modular has a certain kind of value.
• 44:33
  But if you figure out how to do the operator thing, the value
• 44:36
  just goes through the roof.
• 44:40
  And whether you claim it or not, or capture it or not uh, someone's
• 44:44
  gonna capture it and figure out how to turn it into a, a real business.
• 44:48
  And that's, that's where Clay, you know, Clay's work was really about how incumbent
• 44:54
  firms have a hard time reallocating resources into the new stuff, 'cause the
• 45:00
  new stuff in his disruption world is sort of not as good as the old stuff, right?
• 45:07
  A lot of what we see in modularity and, and its working out is different.
• 45:14
  In that case, you have incumbent firms that are big enough and
• 45:20
  see clearly the threat, and they do allocate the resources.
• 45:25
  They do put energy and work into creating the new stuff, and they
• 45:29
  actually create the new stuff, but it doesn't work as well.
• 45:34
  They have a hard time making it work as well because they don't understand all
• 45:39
  the nuances and all the, as Carliss said, all the off-diagonal little connections
• 45:47
  that, are in the new technology because they don't have experience with it.
• 45:53
  GM has a hard time making small cars.
• 45:55
  Takes them years, years, to make a really good small car.
• 46:00
  They eventually do, so they survived, but it took them probably 20 years
• 46:06
  before they made a really good small car.
• 46:09
  Took Xerox, never made… They made small compu- small pocket copiers, but
• 46:16
  they never successfully transitioned out in that world 'cause they were
• 46:21
  so encumbered by their, the way they were organized and stuff.
• 46:26
  It was really hard for them to move.
• 46:28
  And there are zillions of other examples like that where people
• 46:31
  actually do t- do try to make it.
• 46:34
  This happened in photolithography.
• 46:36
  the incumbents made the new equipment.
• 46:39
  They all made the new stuff because it wasn't that complicated.
• 46:42
  It was just, you know, "Oh, we gotta figure out how to make this
• 46:45
  work a little bit differently."
• 46:47
  And they all did it, but it didn't work very well because they didn't
• 46:50
  understand all the nuances, it was the nuances that made a difference.
• 46:54
  And their customers could move.
• 46:55
  There was noth- they weren't locked in.
• 46:57
  There was nothing that ke- kept the customers from moving.
• 47:01
  So, whether you exploit the value or not depends on how well you understand it
• 47:05
  uh, that, that technical dynamic as well as what's happening to your customers
• 47:09
  so let me bring this to AI.
• 47:12
  And the other amazing thing in the Henderson-Clark paper is a sentence at
• 47:20
  the beginning of the conclusion that says, " Organizations mirror the structure
• 47:27
  of their tasks." More or less that.
• 47:31
  Which is saying, you know-- I mean, it's, it's so o-- if, if you skip
• 47:36
  a task or you skip a transfer, that's often fatal to the outcome.
• 47:41
  But no one had ever said it before.
• 47:45
  And in order to mirror, you have to understand the
• 47:50
  process at a very micro level.
• 47:54
  why, that's why the DSMs were so fabulous, because they pointed out, at that
• 48:01
  time in management theory and practice
• 48:08
  top managers had abstracted away from processes.
• 48:14
  I, I, I mean, I'm gonna be what- whatever, uh
• 48:20
  Um, disrespectful.
• 48:25
  So strategy came along and said, you know, the, the unit of analysis is
• 48:31
  something akin to a, a company in an army.
• 48:38
  You know?
• 48:38
  The, the top brass doesn't have to know what's going on with the sergeants.
• 48:43
  The sergeants do that.
• 48:46
  And so we, we, you know, we, we think like generals and everybody want--
• 48:53
  everybody wants to be a general, and especially Harvard MBAs wanna be and then
• 49:00
  generals and then commanders-in-chief.
• 49:04
  Well, if, if it's the fine structure that is to see, that is the true basis
• 49:15
  of performance, then you wanna have an understanding of fine structure.
• 49:22
  Today, we're hearing about this in AI The buzzword for
• 49:28
  the DSM right now is workflows.
• 49:31
  Companies, in order to use AI effectively, are having to map their
• 49:36
  workflows, which they didn't do before.
• 49:41
  But now they have to because the AIs need to-- need, need
• 49:48
  help understanding workflows.
• 49:52
  But There's also differentiation across the
• 49:58
  across the steps in the AI process.
• 50:01
  So what gets all the attention are the LLMs, the, frontier models,
• 50:06
  I think they sometimes call them.
• 50:08
  And they are trained
• 50:12
  based on, you know, e-everything ever written by man or woman or child.
• 50:21
  But
• 50:22
  nobody
• 50:22
  sees what goes on inside the frontier model to put the sentences together.
• 50:30
  We don't-- That's very much a black box.
• 50:33
  Anthropic has tried one thing where they can, they can isolate certain words to
• 50:38
  certain sections of the LLM, like people can see what parts of your brain are
• 50:43
  activated by certain triggers, thoughts.
• 50:49
  But nobody knows how the LLMs are truly wired because they're,
• 50:58
  they don't learn that way.
• 51:00
  Now we have agents, and with agents, we're-- To, to, to access the power of
• 51:07
  agents, you have to think of the agents as modules, and you hook them together in
• 51:13
  an architecture to do different things.
• 51:16
  My, my, my daughter treats her life this way.
• 51:22
  She has her recipe agent, she has her-- she has a whole bunch of agents
• 51:29
  doing different things in her life.
• 51:32
  And then you have to more formal about how the agents interact.
• 51:39
  And I think, now I've not been part of any AI deployment at an actual company,
• 51:45
  but I think where they're struggling in the fact that they have never had cause
• 51:53
  to actually map the tasks and transfers that go inside-- go on inside units
• 52:01
  that are called divisions, departments,
• 52:04
  offices,
• 52:04
  teams, because, you know, it hasn't paid or they don't think it pays.
• 52:12
  But now, because the AI demands it, all struggling to map their workflows
• 52:18
  this is a perfect segue for a great story that you tell and where there's
• 52:25
  a real threat with AI of the black box.
• 52:27
  And I, I'll give it, I give you… 'Cause you told the story
• 52:30
  about the supermarket, Kim.
• 52:32
  I had one of these recently where about a decade or more ago when I
• 52:37
  moved into the house I'm now living in, actually it's, it's going on
• 52:40
  13 years, I planted a hedgerow around the border of my just garden.
• 52:46
  And th- there used to be a road that ran through it, so a certain
• 52:51
  part of the hedge has literally got compacted soil underneath, and I
• 52:55
  took out three skips worth, three big massive bins worth of soil.
• 53:00
  I changed that soil.
• 53:01
  But this piece just of the hedge struggled to grow.
• 53:05
  But we had a landscaper who land, who comes in and cuts all the hedges and the
• 53:08
  gardens in where I live, and he, he knew.
• 53:12
  I had told him, was like, "Okay, this, just understand
• 53:15
  here, this runs under here.
• 53:16
  This piece is struggling to grow.
• 53:18
  We need to have patience with it.
• 53:19
  I'm feeding it." So he knew about it.
• 53:22
  Then the management company who looks after our grounds changed the landscaper.
• 53:29
  A new landscaper comes in.
• 53:30
  I come home one day.
• 53:32
  T- it was only recently.
• 53:34
  Butchered my entire fence, my hedge.
• 53:37
  Like years of growth gone, and he took it down to the lowest common
• 53:41
  denominator and thought I, he was doing me a favor 'cause it looked neat.
• 53:45
  And I thought about your study of the automobile industry and how institutional
• 53:51
  knowledge can be lost and has dramatic effects, way worse than my hedge,
• 53:57
  as you showed with the automobile and the hood, and the oscillation
• 54:01
  and the vibrations of the hood.
• 54:03
  I'd love you to tell this story because it absolutely exemplifies what Carliss
• 54:07
  was saying about the threat of the black box and the threat of losing
• 54:11
  knowledge, and actually know-how
• 54:13
  this was a, this was a real problem for the automobile companies in the US who
• 54:19
  were moving from big cars to small cars.
• 54:23
  And the hood of the vehicle is a, a good example of the problems they faced, right?
• 54:29
  Because the hood covers the engine compartment, and in a
• 54:32
  large vehicle, there's a lot of room in the engine compartment.
• 54:37
  And so they learned over years, and it really was tacit knowledge.
• 54:43
  It wasn't it wasn't formally, you know, written down anywhere.
• 54:48
  It wasn't part of a modularization.
• 54:51
  It was just guys figured out how to do it, you figured out where to put
• 54:55
  stuff and what the dynamics of the vehicle were, its, its noise, vibration,
• 55:03
  harshness is the measure they use.
• 55:06
  And they had these rules about how you do this.
• 55:08
  When you do a small car, the hood is much smaller, the engine
• 55:11
  compartment is much smaller, and there's not much room in there.
• 55:15
  And so you have to figure out how to, quote, "package" the engine compartment
• 55:21
  so that the vibration of the vehicle and the acoustics especially, work
• 55:26
  in such a way that it's a pleasant experience for people who drive, right?
• 55:31
  So the first small cars they introduced were subject to a lot of vibration
• 55:38
  that they were not familiar with.
• 55:41
  And so literally, when they first took these cars out on to test
• 55:45
  them, the hoods would vibrate,
• 55:47
  Oh, geez.
• 55:47
  Oh
• 55:48
  and they would, they would, they literally would, would roll
• 55:54
  dear
• 55:54
  a wave because there wa- there literally were waves wa- waving through the steel.
• 56:00
  The steel's very thin in a hood, if you ever go look at it.
• 56:04
  It's very thin, and it was, it was waving because of the acoustics and
• 56:10
  the vibration that they created.
• 56:13
  And it was very difficult for them to figure out how to make
• 56:17
  a small engine compartment and a hood that didn't vibrate.
• 56:22
  They had to kind of create a whole new set of, quote, "design rules" that you put…
• 56:28
  And 'cause essentially the engine compartment was very modular, It's,
• 56:33
  it's got certain things that have to be in there, and where you put them
• 56:36
  and how they hook together is you know, these are all modules and so forth.
• 56:40
  it, it goes down to things like, you know, how, how you build an engine
• 56:45
  what the engine looks like, and how you build, you know, the battery.
• 56:49
  Where do you put the battery?
• 56:50
  And, you know, you've got a, you've got a coolant system, and where do
• 56:53
  you put that, and how do you put the-- where do you put the pipes
• 56:56
  and the tubes and all that stuff?
• 56:58
  Anyway
• 57:00
  the, the best example, worst example of of the problems that this created
• 57:06
  was the a, a vehicle called the, the Vega, and it was created by Chevrolet.
• 57:13
  And I happen to own one, unfortunately.
• 57:17
  But when the Vega was first put on the test track to drive around,
• 57:24
  it liter- literally fell apart
• 57:29
  which meant that the vibration in the vehicle so severe that it actually tore
• 57:36
  some of the, the, know, the fasteners and the bolts and stuff, they sheared
• 57:42
  off and the whole thing-- Literally, the thing fell apart on the test track.
• 57:46
  Oh my
• 57:48
  And they realized they had a whole different animal
• 57:51
  that they were dealing with.
• 57:53
  And-- there was-- It was very hard.
• 57:56
  It was really, really hard to figure out how to build a
• 58:02
  really, really good small car.
• 58:05
  And one that not just wouldn't fall apart, but it would be really
• 58:08
  smooth and, you know, low cost and no harshness in when you drive
• 58:17
  it and, you know, all that stuff.
• 58:19
  You know, it was a, you know… it's exactly what we've been talking about.
• 58:24
  They got, they got hit by the little off diagonal X's that they
• 58:29
  hadn't, they hadn't-- they didn't even know existed, you know?
• 58:34
  And they had-- And it just took them a long, long time.
• 58:37
  And they tried, and they, they got better, and they got better and better.
• 58:40
  And today, the American companies have actually abandoned small cars pretty
• 58:46
  much because they don't make money.
• 58:49
  And in fact, I think Ford doesn't make sedans anymore at all.
• 58:52
  They just make SUVs and trucks and and so forth.
• 58:58
  So fascinating
• 59:00
  there's a really important lesson in there that you could easily miss.
• 59:04
  So, so many people in organizations, people who have great ideas,
• 59:09
  w- may articulate those ideas, and they don't go anywhere, and
• 59:12
  they get very, very frustrated.
• 59:14
  You give an example of a kid called Jake and my son's name's
• 59:17
  Jake, so it resonated with me.
• 59:19
  But you have an idea, l- let's go with the mug or the, the hood, and I'm like going,
• 59:24
  "Why don't we use a lighter material?" like goes and spends ages trying to design
• 59:29
  this, maybe create a prototype, present it to the company, and it totally bombs.
• 59:34
  But as you say, and, and the huge thesis in the book is you
• 59:39
  can't just change one module.
• 59:41
  It affects the entire thing.
• 59:43
  And I, I think that I'd love you to share that as a concept that our audience can
• 59:49
  really grasp, because you can see where frustration comes from if you don't
• 59:54
  understand how the modules stick together
• 59:57
  Um, well, the optionality comes from the fact that there's some things you
• 1:00:06
  don't have to track and that's the trick of reducing the interactions
• 1:00:13
  from the mess that is the natural state of a complex process to a bunch
• 1:00:22
  of pinch points coordinated by design rules and by system integration.
• 1:00:29
  That's a big learning process.
• 1:00:32
  And to do it consciously, tr-truly it, it was an invention, not an evolution.
• 1:00:43
  was not where the organizations naturally wanted to go.
• 1:00:48
  Divisional corporations are a natural evolution out of smaller corporations.
• 1:00:56
  You hook them together, you create a staff that understands or, is
• 1:01:02
  socialized to provide the glue, and you have kind of a, a dirty modular
• 1:01:09
  structure with blocks then connections between the blocks, and a lot of the
• 1:01:15
  connections are provided by the staff.
• 1:01:20
  But when you have that mindset
• 1:01:26
  there are places you can't go.
• 1:01:28
  Like GE tried to do computers, and GE was a behemoth in the '50s and '60s.
• 1:01:38
  I mean, they were-- they, they did everything.
• 1:01:42
  General Electric, Thomas Edison's company.
• 1:01:46
  that used electricity that was in the purview of General Electric.
• 1:01:54
  And but they tried to do computers, and it never worked for them because .. it's very
• 1:02:04
  natural for a big company to be modular.
• 1:02:08
  The danger is that the modules fly off and, you know, you need to think very hard
• 1:02:14
  about where the points of control are that the value doesn't all flow to the modules.
• 1:02:20
  They have to remain beholden to the platform in some way.
• 1:02:26
  But,
• 1:02:29
  If you start out as a line staff company and you wanna become modular,
• 1:02:38
  there's nothing in your DNA, in your corporate DNA, that would
• 1:02:42
  suggest to you what IBM did.
• 1:02:46
  I mean, it, it took IBM-- had to set up
• 1:02:52
  many new centers be very strict to prevent the modules from depending on one another.
• 1:03:04
  And they were es-- almost the entire rest of the industry was de novo computers.
• 1:03:12
  They didn't have a past like IBM.
• 1:03:16
  They… and so as with Amazon and online
• 1:03:24
  they weren't mired or constrained by a past set of practices.
• 1:03:33
  If you look at, you know, all the generations
• 1:03:40
  after IBM, Kim was mentioning Intel, Microsoft.
• 1:03:44
  I mean, these are modular native companies that started small and got big.
• 1:03:55
  So, so that there really is,
• 1:04:01
  Hi- history constrains you.
• 1:04:02
  Your history constrains you a lot.
• 1:04:06
  Just as I'm hopeless at AI.
• 1:04:08
  I-- So not AI native.
• 1:04:12
  I'm trying, I really am.
• 1:04:15
  but I'm not AI native.
• 1:04:17
  And I just wanted to put in one thing about design rules before there were
• 1:04:25
  des- instructions before the d- there were design rules, were in the form of
• 1:04:30
  switches in a large array of circuits.
• 1:04:36
  A array bigger-- Or the array was as big as a room.
• 1:04:41
  And you had women, usually young women, who went and flipped the
• 1:04:49
  switches to make the instructions, to make the circuitry correct.
• 1:04:55
  And there came to be a term, called a kilogirl, and it was what you needed to do
• 1:05:07
  1,000 instructions per hour, a kilogirl.
• 1:05:13
  Only I think it was a day.
• 1:05:16
  But, it became a standard.
• 1:05:18
  A standard was a kilogirl.
• 1:05:22
  So that, standard went away, but it was there for a while.
• 1:05:29
  And those were the days when a big company would have rooms full of
• 1:05:34
  women on calculators putting in the f- transactions for the company.
• 1:05:44
  The processing speed of one of those women would be the measure of a kilogirl
• 1:05:52
  So,
• 1:05:52
  you know, let me, let me just say a word about what Carliss has introduced there.
• 1:05:56
  Not about the Kilo girls, but about,
• 1:05:59
  about the difficulty of, of, as you said, of a module coming into being that might
• 1:06:06
  o- it actually might obey the design rules, but it also has other interactions
• 1:06:10
  that have not been anticipated, right?
• 1:06:14
  So imagine the following.
• 1:06:19
  Imagine an organization that is consciously modular itself.
• 1:06:29
  This is another thing that we, we wrote a paper in 1997 or published
• 1:06:34
  in 1997 in Harvard Business Review.
• 1:06:38
  And if you go back and read that paper, the last two paragraphs,
• 1:06:43
  I think, lay out the possibility.
• 1:06:45
  This is, this is a apropos of what Carliss called the mirroring hypothesis we set
• 1:06:51
  out and we said, "It's possible that modularity could become a feature of the
• 1:07:00
  organization itself, as the organization could be consciously, explicitly
• 1:07:08
  modular." So imagine a situation where you have that, and, there are a couple
• 1:07:15
  of things, there are a couple of things that have to be in place for modularity
• 1:07:22
  to work at all, and are part of the-- Design rules are part of it, but you
• 1:07:27
  also have to have an architecture figures out, well, what are the modules?
• 1:07:33
  And s- and thirdly, you have to have a way to test a module that it will
• 1:07:40
  actually function well in the system before… So you don't have to design and
• 1:07:45
  put, put in operation the whole system.
• 1:07:47
  You can test the module independently.
• 1:07:49
  So imagine you've got that and you've decided, "Well, we're gonna try to create
• 1:07:53
  an organization that's sort of modular." And so we've written about this in this
• 1:08:00
  book we published a couple years ago.
• 1:08:03
  And what we decided as we worked on that was that, and we've-- and even
• 1:08:10
  subsequent work that we've done, is that it turns out that leadership
• 1:08:17
  is actually the operating system the organization.
• 1:08:22
  And it's, it's the way decisions get made, it's the way information moves
• 1:08:27
  around, it's the way people interact with one another and so forth.
• 1:08:33
  And if you design a company-- You could design a company that is
• 1:08:37
  actually modular itself around teams.
• 1:08:41
  And if you create a setup where the leadership becomes the operating
• 1:08:48
  system for a modular system, okay?
• 1:08:52
  So it's gotta have that character.
• 1:08:54
  Then you can actually do it, and there are some examples.
• 1:08:58
  So one example you may be familiar with is Haier.
• 1:09:02
  Mm-hmm.
• 1:09:03
  It's a Chinese appliance company.
• 1:09:05
  I don't know if you're familiar with what happens-- has happened at
• 1:09:07
  Haier in the last 10 years or so.
• 1:09:10
  But Haier went from having a classic fully integrated organization with
• 1:09:19
  12 different layers of management.
• 1:09:23
  It was a f- it was like a, at the time, it was probably like a $30 billion
• 1:09:28
  global organization headquartered in China, but operating all over the world.
• 1:09:34
  Haier, by the way, apropos of what Carla said, bought GE Appliances.
• 1:09:40
  And but a few years ago, Haier decided that they were on-- that they were in
• 1:09:46
  a death spiral They couldn't move fast enough with their big, clunky organization
• 1:09:53
  to respond to technology changes and competitive moves and global dynamics.
• 1:09:58
  They needed a much, much different organization.
• 1:10:01
  And, and bless their hearts, they completely created a
• 1:10:07
  complete different organization.
• 1:10:08
  It took them a while to do it.
• 1:10:10
  And what they did is they, they created, at the bottom,
• 1:10:16
  their modules, which are teams.
• 1:10:20
  And there are-- in hire today, there are 4,000 of them.
• 1:10:24
  And above them is one layer, and it's called, it's called ecosystems,
• 1:10:31
  These are… So the teams are all in different things like, I mean,
• 1:10:36
  one team is actually a factory.
• 1:10:39
  So the GE appliance factory in Kentucky is, in Louisville is, is a team.
• 1:10:44
  Anyway, they got all these teams and, and they have designed it so that
• 1:10:49
  the staff is where the resources are, but the action is all in the teams.
• 1:10:58
  And the senior executives no longer instructions or commands.
• 1:11:05
  There's noth- it doesn't work that way.
• 1:11:07
  These senior executives now are investors and architects.
• 1:11:15
  That's what they do.
• 1:11:16
  They architect and they invest.
• 1:11:19
  then the middle layer, there's only three layers.
• 1:11:22
  middle layer are people who watch over the ecosystems.
• 1:11:27
  And then you have an ecos-ecosystem leader, and then you've got… And
• 1:11:30
  ecosystem leaders basically are-- it's a very modular structure, meaning
• 1:11:37
  that the teams can, they can kinda come up with whatever they want to.
• 1:11:43
  They can get resources internally.
• 1:11:45
  They have to contract with the staff for resources, or they can
• 1:11:51
  go outside if it's a better deal.
• 1:11:54
  And so they become very entrepreneurial.
• 1:11:56
  But in order to function that way, there has to be information
• 1:12:01
  for-- about the architecture.
• 1:12:03
  And the architecture for that kind of organization includes: what's
• 1:12:06
  our strategy What are our values?
• 1:12:11
  What are the principles that we want our company to operate under?
• 1:12:17
  What are our objectives specifically?
• 1:12:19
  And everybody, all those 4,000 teams and the, you know, the
• 1:12:24
  teams are like 10 to 15 people.
• 1:12:27
  So you've got thousands of people and, you know, they have a big organization.
• 1:12:33
  Say they're like… I think they're almost $70 billion in revenue.
• 1:12:37
  Anyway, here's an example of a team.
• 1:12:42
  There was a team whose responsibility was rural washing machines, or washing
• 1:12:49
  machines sold in rural areas of China.
• 1:12:52
  Okay, so this is-- it's in, it's in China, and they're selling
• 1:12:55
  washing machines to rural customers.
• 1:12:57
  That's the team, right?
• 1:12:59
  There's like 10 or 15.
• 1:12:59
  They design the products.
• 1:13:02
  figure out where they're gonna be made.
• 1:13:04
  They contract with the, you know, the operations about how
• 1:13:06
  to distribute it and market.
• 1:13:07
  So they get it, they build all that, right?
• 1:13:09
  They're the team.
• 1:13:11
  They discovered something, and that is that their customers were using
• 1:13:16
  washing machines to wash vegetables
• 1:13:22
  And they'd put the vegetables in the washing machine.
• 1:13:24
  They wouldn't put soap in, they'd just turn it on and, and wa- go through
• 1:13:28
  the wash cycle and, and it turned out it worked really well, right?
• 1:13:32
  So the guys thought, " Well, why don't we do that and make that
• 1:13:36
  a feature of the machine so we can do it a whole lot better?
• 1:13:41
  'Cause we can create better baskets, we can create a s- an actual cycle that
• 1:13:45
  does that." And they designed it, they worked with engineering to get it done.
• 1:13:50
  They figured out how to make it, get the products made, and they s- it became
• 1:13:53
  like, I think it was like $100 million business, because it was a really
• 1:14:00
  valuable thing for their customers.
• 1:14:03
  old Haier would never have figured that out.
• 1:14:07
  They couldn't possibly have figured it out.
• 1:14:09
  So the idea is you put this team really close to customers, you
• 1:14:13
  give them flexibility, they have a framework they operate under,
• 1:14:19
  which is: what's our strategy?
• 1:14:21
  are our values?
• 1:14:22
  What principles do we use?
• 1:14:23
  What are our objectives?
• 1:14:25
  So that kind of very visible information goes way down.
• 1:14:30
  So somebody like, you know, operating in whatever country knows as much about
• 1:14:36
  the strategy of Haier as the CEO, and then they can make good decisions, and
• 1:14:40
  the decision work together, and they figure out how to make… And, you know,
• 1:14:43
  the staff is a support group, not…
• 1:14:46
  They don't tell people what to do.
• 1:14:48
  So it's possible to do this.
• 1:14:50
  It's possible to create.
• 1:14:51
  And they're coming up with modules in these teams that nobody
• 1:14:56
  ever thought of before, right?
• 1:14:57
  But because the organization is explicitly modular, anticipate that that will happen.
• 1:15:05
  It will happen, and it's a good thing, because now we figure out,
• 1:15:08
  "Oh, how do we make this work?"
• 1:15:10
  And that says, "Okay, what are…" And Carliss is, she loves DSMs.
• 1:15:14
  What are the Xs, right?
• 1:15:16
  No
• 1:15:17
  this gonna work?
• 1:15:18
  What are the Xs?
• 1:15:18
  How do we build it?
• 1:15:20
  How do we connect it?
• 1:15:21
  And they work with the staff and, you know.
• 1:15:23
  But it's happening 4,000 teams across the whole world.
• 1:15:26
  Haier has grown much faster.
• 1:15:29
  It's much more nimble.
• 1:15:30
  It's much lower cost than ever before, and it's like a dynamo
• 1:15:36
  because they've unleashed the power of modularity in the organization
• 1:15:41
  Di- did you write about that in Leading Through?
• 1:15:41
  'Cause I, uh, that's where I read it from.
• 1:15:42
  I, I love telling the
• 1:15:43
  Oh, did we?
• 1:15:43
  No, did we?
• 1:15:44
  Yeah, we, I think we had one little example in there.
• 1:15:47
  certainly,
• 1:15:48
  we certainly have done it, we certainly have done it in the
• 1:15:51
  work we've done subsequently.
• 1:15:53
  Yeah, they, well, it explains to me how you understand how you've got to your
• 1:15:57
  level of thinking now, and like Carliss said earlier on, IBM was the environment.
• 1:16:02
  And to me, it's like where you're working now with the Leading Through
• 1:16:05
  Institute, it's like leadership is the environment to enable this now.
• 1:16:10
  It's the same, it's the same thing, and if you understand the le- the lens
• 1:16:14
  of modularity, you can arch- architect the organization to be modular.
• 1:16:18
  The final… I we need to finish up now, but I, the thing I really got
• 1:16:22
  out of the book was this as well.
• 1:16:24
  I mean, so many aha moments, and I, I love going back and really getting
• 1:16:29
  my hands dirty with this content.
• 1:16:31
  But 1968 Conway, Conway's Law that the, the product is, is imprinted in
• 1:16:39
  the organization structure, and then your paper, Kim, with Rebecca in 1990,
• 1:16:45
  that actually, that, that's the mirror.
• 1:16:47
  It works both ways.
• 1:16:48
  It actually gets back, imprinted back into the product as well.
• 1:16:52
  And then both you and Carliss, Carliss Clark that, or Kim, your, your content
• 1:16:59
  together lets you understand this.
• 1:17:01
  And then we go, "Well, actually we can do something about it." And that's,
• 1:17:05
  that was kind of my, one of the many, many moments of, of joy of understanding
• 1:17:10
  this content that came to me as well.
• 1:17:12
  So it's it's, it's a brilliant book.
• 1:17:15
  And you know what?
• 1:17:16
  We didn't even get near, near any of the things like substitution
• 1:17:20
  and all the different kind of ideas that are splitting that are in the
• 1:17:25
  book, all these great concepts.
• 1:17:26
  Maybe we'll do a part two and sh- share those six, the, the six real key
• 1:17:32
  ideas that you can actually ar- artif- ar- architect a modular organization
• 1:17:39
  if you understand these things.
• 1:17:41
  Kim, I'll come to you first and, and maybe share your, your final thoughts,
• 1:17:45
  but also the work you're doing today and where people can find you.
• 1:17:49
  well, first of all, thank you for putting us together.
• 1:17:53
  Carliss and I started working together about 1985 or '6,
• 1:17:58
  I think, somewhere in there.
• 1:17:59
  Yeah
• 1:18:00
  And we worked together for, you know, 20 years.
• 1:18:05
  Yeah
• 1:18:07
  and this is exactly what it was like, back and forth, trying…
• 1:18:14
  We had a lot of fun together.
• 1:18:16
  So the thank you for putting us together.
• 1:18:19
  I will just say that, It's a little bit like the little boy with a hammer.
• 1:18:25
  Everything looks like a nail.
• 1:18:26
  And once you understand modularity, it's, everywhere.
• 1:18:32
  And it's really, really powerful.
• 1:18:34
  And Carliss' explication of agents, AI agents and workflows and stuff
• 1:18:40
  is a, is really a wonderful example of how this concept can help you
• 1:18:45
  understand things that otherwise you have no idea what's going on.
• 1:18:50
  So it's a really powerful idea, and people have understood that now, and it's,
• 1:18:55
  it's a commonplace, but it was a real joy to work on, on this stuff together
• 1:19:01
  and to discover these things and to see them in action is is really wonderful.
• 1:19:07
  And I think what we're doing now, we published this book a couple of years
• 1:19:12
  ago called "Leading Through: Activating the Soul-Heart-Mind of Leadership."
• 1:19:17
  And what we've learned since then is to think of leadership as the operating
• 1:19:22
  system of the organization, analogous to an operating system in a computer
• 1:19:29
  where leadership really does carry that weight organization.
• 1:19:36
  And every organization-- in, in every organization, leadership
• 1:19:40
  is the operating system.
• 1:19:42
  And some operating systems work really well, and some are horrible,
• 1:19:46
  and they, they chew people up.
• 1:19:49
  They they fail their moral responsibility.
• 1:19:53
  They actually hamper potential in the organization and so forth.
• 1:19:59
  Others are much more, powerful because they become… The closer
• 1:20:04
  they get to becoming truly modular.
• 1:20:07
  So that's why we call it the modular leadership system, is
• 1:20:10
  that leadership is not just… Is, is it is definitely personal.
• 1:20:14
  It's a really important part of it, but it's also organizational.
• 1:20:18
  So that's an important thing we're working on.
• 1:20:20
  We've created this thing called the Leading Through Institute, and we've
• 1:20:23
  now, we've now come to understand a deeper thing, which is that the, the
• 1:20:31
  leadership, the way we defined it, coupled with AI, because we've now
• 1:20:36
  created a learning platform we call Permios, which is allows people to
• 1:20:41
  learn the principles of leadership and the operating system really deeply.
• 1:20:46
  So it's not just what they know.
• 1:20:48
  They re- they understand its value, they feel it, they do things,
• 1:20:52
  they become different people.
• 1:20:55
  And in that context the combination of modularity coupled with this AI
• 1:21:01
  system means that what you create is you create a, the infrastructure for
• 1:21:09
  any business system That you want to implement in order to operate.
• 1:21:16
  and it becomes the, the, the structure that you use.
• 1:21:18
  It's sort of this underlying structure, but it's inherently modular.
• 1:21:22
  So you design it that way.
• 1:21:24
  design the AI platform that way.
• 1:21:27
  So the platform itself is modular, and it's, it, and it's, I mean,
• 1:21:34
  Aidan, it's just like incredibly
• 1:21:35
  powerful
• 1:21:36
  Unbelievably powerful.
• 1:21:38
  And I'll link to your website as well, Kim.
• 1:21:41
  The, it's leadingthrough.co.
• 1:21:43
  But I'll link to that and I'll link to the episode we recorded
• 1:21:46
  with your brilliant daughter and your son, John and Erin as well.
• 1:21:50
  I'll share that in the show notes as well.
• 1:21:53
  But Carliss, coming to you, I learned so, so much from you.
• 1:21:55
  I wanted to ask you because on the, your second volume Design Rules Part Two, the
• 1:22:02
  cover, the cover has a, a signification, and I wanted to, I wondered what it was
• 1:22:07
  yes.
• 1:22:09
  Let me
• 1:22:10
  Kim has a copy there, Carliss.
• 1:22:11
  Look.
• 1:22:13
  Ki- Kim has it there, Carliss
• 1:22:16
  Yes.
• 1:22:19
  So that is a picture I took of the inside of St. Mark's in Venice.
• 1:22:27
  I own the copyright to the picture.
• 1:22:30
  There is a famous essay Stephen Jay Gould and Richard Lewontin
• 1:22:38
  "The Spandrels of San Marco."
• 1:22:43
  and the ar- the, the, the spandrels are the, you know, the arch goes
• 1:22:49
  up and then the, the space between
• 1:22:54
  bones of the arch, are the spandrels.
• 1:22:58
  So you can see a couple of spandrels in this picture the, in, in the middle.
• 1:23:05
  Gould and Lewontin's argument was one I found compelling.
• 1:23:13
  It was that
• 1:23:17
  They, they were undergoing some kind of debate with their colleagues about
• 1:23:22
  evolution, and they said, "You know, some things exist because the physics
• 1:23:29
  requires it." So the spandrels do not-- The spandrels exist because if you want
• 1:23:36
  an arch of that kind, you need that space.
• 1:23:40
  You need those quasi-triangular areas.
• 1:23:44
  It was not that somebody said, "I wanna create art in quasi-triangular
• 1:23:50
  areas," although all the spandrels are filled in with beautiful art.
• 1:23:56
  it's the, the architecture demanded that the space be there, and then
• 1:24:03
  creative people made use of that space.
• 1:24:07
  And thing-- people are more respectful of technology now.
• 1:24:13
  Kim was, Kim was a huge influence on management and, and on all of us
• 1:24:21
  about the importance of technology.
• 1:24:24
  But there's a, there's a whole school scholars that don't think
• 1:24:32
  technology matters, you know?
• 1:24:34
  You can, you can have anything you want, in, in an organization.
• 1:24:39
  It's, it's all socially constructed anyway.
• 1:24:44
  And I always felt, you know, that, that since… since…
• 1:24:49
  recruited me after I got tenure.
• 1:24:53
  I was in the finance group, and I didn't have any idea what I was gonna do next.
• 1:24:58
  And he said, "Come and learn about technology. Not about production, but
• 1:25:03
  about technology." And it was, it was-- say it was eye-opening, it was, it
• 1:25:11
  was a, a, an amazing world unfolded.
• 1:25:16
  But one of, one of the things that,
• 1:25:21
  that became a conviction, was that the world will be what it is,
• 1:25:27
  whether you like it that way or not.
• 1:25:30
  And you just can't argue physics or chemistry or bi-
• 1:25:35
  you know, you, you just can't.
• 1:25:37
  And so technol- that, that made me feel that technology is the foundation of
• 1:25:44
  management, because that's, that's, that's where the rubber meets the
• 1:25:48
  road, and just gotta deal with it.
• 1:25:52
  You can't wave your hand and, and say, "Oh, I'd like an organization to
• 1:25:57
  be this way," without comprehending that there's some rules.
• 1:26:02
  A- anyway, so, so the book's cover, I was looking for a cover and I
• 1:26:10
  said, "You know what ex-" And I-- there's nothing in the book that even
• 1:26:15
  mentions the spandrels of San Marco.
• 1:26:18
  this is the backstory to that cover, which I like very much
• 1:26:23
  Beautiful.
• 1:26:23
  Well, you two are the spandrels of
• 1:26:27
  Right
• 1:26:28
  this n- of this content, and it's been a, an absolute
• 1:26:31
  pleasure having you on the show.
• 1:26:33
  I'll link also, Carliss, to where people can buy the book as well, and
• 1:26:36
  part one as well of the the book.
• 1:26:42
  See this, but the headline, the tagline is, is the key, how
• 1:26:46
  technology shapes organizations
• 1:26:48
  Yeah.
• 1:26:48
  And, and, and that's, you know, it's not how organizations shape technology.
• 1:26:52
  It's about, you know, you deal with the technology and your organization
• 1:26:58
  follows, which is probably not exactly… Well, so, so K- Kim is
• 1:27:18
  the leadership expert.
• 1:27:20
  I've never purported to be an expert, although I have to say Kim's leadership
• 1:27:23
  is a lesson in itself, and what he did at the school was amazing.
• 1:27:28
  What he's doing with his institute is amazing.
• 1:27:32
  Taught, me a whole lot.
• 1:27:34
  Taught me everything I know.
• 1:27:36
  But I never,
• 1:27:41
  It was never my ambition to be a leader.
• 1:27:46
  I, I never thought it was something I could aspire to.
• 1:27:50
  So anyway, that's, that's neither here nor
• 1:27:52
  there.
• 1:27:53
  This book nails this topic.
• 1:27:56
  absolutely nails it.
• 1:27:58
  really, really hard to argue with this book.
• 1:28:02
  And that has, that has a profound influence because so much of today's
• 1:28:07
  society is based on something completely different, which is there isn't,
• 1:28:12
  there is-- are no spandrel effects.
• 1:28:16
  That is that everything is… There's no such thing as truth.
• 1:28:21
  You create your own.
• 1:28:23
  no, there are no absolute laws.
• 1:28:26
  There's nothing that, you know, you just kind of create your own.
• 1:28:29
  And everything's relative, and that turns out not to be true.
• 1:28:36
  Amen.
• 1:28:36
  Amen.
• 1:28:37
  The guys, it's been, it's been a huge privilege, and I have to say to you,
• 1:28:41
  Carliss, y- you said earlier on that you didn't know anything about AI AI, or now
• 1:28:45
  you're saying about leadership as well.
• 1:28:47
  You sound a little bit like IBM when they had Intel in their grasp.
• 1:28:50
  You have the ingredients, and you've proven it.
• 1:28:54
  It's been an absolute pleasure.
• 1:28:56
  Authors of Design Rules, Kim Clark and Carliss Baldwin, thank you for joining us
• 1:29:01
  Thank you
• 1:29:02
  Thank you.